Atlas proposal

[Samuel Thibault]

Roger Leigh, le Tue 17 Aug 2010 22:45:50 +0100, a écrit :
> Why can't this be fixed the correct way:
> by building all optimised variants for a given architecture and
> selecting the appropriate variant at runtime based upon the system's
> capabilities e.g. from CPUID on i386/amd64?

Because atlas doesn't optimise only for the instruction set, but also
the number of available cores, the size of the caches, etc. etc.  Ask
anybody who has done serious benchmarks with atlas, he will tell you
that providing an "optimized" binary is hopeless.

> Disabling threading is also suspect: how can the optimal number of
> threads possibly be determined at build time?

Because it changes how Atlas will statically schedule the computation
kernels.

> In short, Atlas' approach to optimisation by detecting everything at
> build time is wrong.

It's not "wrong", it's HPC.  And HPC people will happily rebuild the
package to get an optimized version.

> it should be actually fixed, probably best if done in collaboration
> with upstream.

Upstream will not go the dynamic way.

Samuel